Artificial Intelligence (GPT-RAG)

Dependent Azure Resource Recommendations

Recommendation	Provider Namespace	Resource Type
Monitor Batch account quota	Batch	batchAccounts
Create an Azure Batch pool across Availability Zones	Batch	batchAccounts
Deploy a PAYG instance of the model with provisioned throughput to manage overflow effectively	CognitiveServices	accounts
Ensure that models are deployed using Global batch for large scale processing	CognitiveServices	accounts
Ensure AOAI models are deployed using Data Zone Standard for data residency requirements	CognitiveServices	accounts
Use comprehensive monitoring solution for AOAI service	CognitiveServices	accounts
Create the Azure machine learning registry in multiple regions	MachineLearningServices	registries
Plan for a multi-regional deployment of Azure Machine Learning and associated resources	MachineLearningServices	workspaces
Deploy Azure Machine learning workspace in secondary region	MachineLearningServices	workspaces
Ensure to create Machine Learning Compute resources in secondary region	MachineLearningServices	workspaces
Configure API management service in multiple Azure regions	ApiManagement	service
Migrate App Service to Availability Zone Support	Web	serverFarms
Deploy Azure Bastion across Availability Zones	Network	bastionHosts
Deploy Azure Bastion into the virtual network in secondary Azure region	Network	bastionHosts

General Workload Guidance

Summary

Recommendation	Impact	Category	Automation Available	In Azure Advisor
Enable Application insights to monitor online endpoints for Azure Machine learning	Medium	Monitoring and Alerting	No	No
Create Application insights for Machine learning workspace in both the regions	Medium	Disaster Recovery	No	No
Use Azure Traffic Manager to route search requests via geo-located clients for failover redundancy.	Medium	High Availability	No	No

Details

Enable Application insights to monitor online endpoints for Azure Machine learning

Impact: Medium Category: Monitoring and Alerting

APRL GUID: 47176883-2602-455b-97f0-6c6a371bf928

Description:

Azure Machine Learning uses integration with Azure Monitor to monitor metrics and logs for online endpoints. Use the built-in features of Application Insights (such as Live metrics, Transaction search, Failures, and Performance) for further analysis.

Potential Benefits:

Monitoring endpoints gives visibility into metrics and logs

Learn More:

Monitor online endpoints using Application Insights

ARG Query:

Click the Azure Resource Graph tab to view the query

Azure Resource Graph

// cannot-be-validated-with-arg

Create Application insights for Machine learning workspace in both the regions

Impact: Medium Category: Disaster Recovery

APRL GUID: 89928e61-bba8-4c25-99c5-a06a09849ecc

Description:

Create Application Insights for machine learning workspace in both regions. To adjust the data-retention period, update and manage data retention in a Log Analytics workspace.

Potential Benefits:

Provides monitoring and data retention in DR region

Learn More:

Data collection, retention, and storage in Application Insights

ARG Query:

Click the Azure Resource Graph tab to view the query

Azure Resource Graph

// cannot-be-validated-with-arg

Use Azure Traffic Manager to route search requests via geo-located clients for failover redundancy.

Impact: Medium Category: High Availability

APRL GUID: a6c8f3f3-199a-4a13-aaf0-eb51eed5352e

Description:

Azure Traffic Manager enhances redundancy by routing search requests to geo-located clients. If a primary search client or region fails, Traffic Manager redirects requests to a backup client in another region. This setup ensures continuous availability and seamless failover for search services.

Potential Benefits:

High availability

Learn More:

Reliability in Azure AI Search

ARG Query:

Click the Azure Resource Graph tab to view the query

Azure Resource Graph

// cannot-be-validated-with-arg